Research

Fertiliser use remains below recommended rates in most of Sub-Saharan Africa, contributing to poor crop yields and poverty. Producers voice suspicion that available fertiliser is often adulterated, but these concerns are (mostly) not backed by reliable evidence. In a prior project with Anna Fairbairn, Hope Michelson, Brenna Ellison and Victor Manyong, summarised on the VoxDev blog,  we visited all fertiliser sellers in Morogoro Region in Tanzania, tested their fertiliser, and found the urea, the most common fertiliser, to be of good quality. However, the beliefs of farmers in the region are pessimistic, likely contributing to lowered demand. 

For this project, I am collaborating with Hope Michelson at the University of Illinois, and Christopher Magomba at Sokoine University of Agriculture to evaluate the effects of an information intervention on the supply and demand decisions of agro-dealers and farmers. We use the random assignment of all 100 markets (in the same Morogoro Region) and nearby villages into treatment and control groups. In advance of the long-rains agricultural season of 2019, we distribute information signs and pamphlets in the treatment markets (as well as inform the agro-dealers), and hold village meetings in the treatment villages. We collected baseline and endline data among both agro-dealers and farmers, and monitored fertiliser sales and prices throughout the agricultural season. A unique feature of our data collection exercise is the inclusion of all agro-dealer shops in the region – allowing us to capture entry and exit into the market.

This research was sponsored by a grant from DFID's Private Enterprise in Low-Income Countries (PEDL) research initiative.

Reducing antibiotics: Evidence from China's Poultry Industry (2018-2020)

China is one of the main poultry producers in the world. The industry is expanding rapidly, but also becoming one of the most polluting in China. In this study, we focus on the use of antibiotics for disease prevention purposes. This use has been associated with an increase of bacterial resistance affecting human health through the food chain, and soil and water pollution.

This study is let by Li Zhou at Nanjing Agricultural University and in collaboration with Meike Wollni at the University of Gottingen. We aim to evaluate the effects of a three-pronged program, developed jointly with a poultry processing company, aiming to reduce antibiotics use among poultry farmers in Jiangsu province in China. The three components include: (i) Raising awareness about antibiotics alternatives, such as improved barn sanitation and the use of acidifiers through the company extension services, (ii) Providing subsidies for acidifiers, and (iii) Exposing farmers to differential social norms. All farmers in our study will receive (i), but components (ii) and (iii) are randomized. We are conducting the baseline survey among 1000 farmers in Spring 2019, and plan to implement a cross-randomized design with 250 farmers in each cell during 2019-2020. The endline survey was completed in 2020.

This research is sponsored by National Natural Sciences Foundation of China.

ISFM-Malawi: Evaluating Malawi's Anchor Farm Program (2014-2019)

The Anchor Farm Model (AFM) of Clinton Development Initiatives (CDI), established in 2008, aims to increase agricultural production, income and food security of smallholder farmers in Malawi through the adoption of yield-enhancing integrated soil fertility management practices (ISFM). To reach this goal: (i) CDI disseminates production knowledge through the use of demonstration plots, farmer clubs, lead farmers and farmer field days; and (ii) CDI offers structured access to credit, seed and output markets. 

In was collaborating with Hope Michelson at the University of Illinois, Chris Barrett at Cornell University, Cheryl Palm at Columbia University,  the late Ephraim Chirwa at the University of Malawi and Wezi Mhango at Bunda College to evaluate the effects of AFM on farmer's adoption, yields and incomes.  We used the random assignment of 250 villages into a control and various treatment groups to establish impacts of the CDI interventions on farmers’ outcomes. Using detailed panel data, we identified the channels through which impacts take place and explored heterogeneity across household composition, underlying risk and time preferences and soil properties.  At the end of the study, we provided the farmers involved with information about their soils and ISFM practices. More information on the former can be found on the 3IE blog Bringing Research Down to Earth

This research was sponsored by the International Initiative for Impact Evaluation.  Data can be downloaded via FIGSHARE. We presented the results of this research at 3IE Delhi  in 2017 (full presentation recorded here and a short interview here) and also discussed the goals, hiccups, and results of our research in a 3IE evidence dialogue webinar in 2020.  The final report of this project can be accessed via the 3IE website here.  For papers following this research, see my publication list on the University's website. A summary of the study on village fairness norms and land rental markets, a spin-off of this project let by Kwabena Krah can be found  on the following site: GlobalDev

Evaluating India's Tara Akshar Literacy Program (2013-2018)

The Tara Akshar + Literacy Program is an adult literacy/numeracy program, designed by Readingwise UK and implemented by the Indian NGO Development Alternatives among illiterate rural women. The program uses an innovative, interactive, group-based, computer-based method program which builds on insights from cognitive psychology and memory tricks. For instance, to teach the alphabet, the shape of the letter is turned into a cartoon that looks like the object that begins with that letter.

I was collaborating with Ashwini Deshpande at the University of Delhi, Alain Desrochers at the University of Ottawa, and Chris Ksoll at Mathematica to evaluate the impacts of Tara Akshar on literacy/numeracy, individual and within-household decision-making and broader socio-economic outcomes. In Phase I (2013-14), we randomly assigned 300 women in control and treatment group and collected data on literacy/numeracy, ability, decision-making (risk aversion, preference consistency, knowledge and confidence). In Phase II (2014-2018), we worked with a sample of 800 women and expanded the data collection to include measures of within-household decision-making (through behavioural experiments) and investment in child health and education. Some results are reported in an Ideas for India Blog 

This research was sponsored by the Social Sciences and Humanities Research Council of Canada. The data of this project can be downloaded via FIGSHARE.  More info can be found on the project site: Tara Akshar Research Project. For papers following this research, see my publication list on the University's website. 

Bt Cotton Adoption in India - among ICRISAT households 2007-2009

For my PhD dissertation at Cornell University, I collected data in collaboration with the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT). The villages selected for this study are part of the Village Level Studies (VLS) program of ICRISAT. In this program, ICRISAT followed 300 randomly selected households from six villages in Maharashtra and Andhra Pradesh during the period 1975-1985. In 2001, ICRISAT restarted the panel, revisiting 185 of the first generation VLS households and their split-offs, in addition to 261 newly added households. 

In this study, I collected information on the determinants of agricultural technology adoption (Bt cotton in particular), with a focus on social networks, among the ICRISAT households in three villages: Aurepalle, Kanzara and Kinkhed.   This research was sponsored by the National Science Foundation and Cornell University in the US.  For papers following this research, see my publication list on the University's website. 

Data can be downloaded via FIGSHARE.

Education in India - among ICRISAT households 2007-2009

For my PhD dissertation at Cornell University, I collected data in collaboration with the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT). The villages selected for this study are part of the Village Level Studies (VLS) program ICRISAT. In this program, ICRISAT followed 300 randomly selected households from six villages in Maharashtra and Andhra Pradesh during the period 1975-1985. In 2001, ICRISAT restarted the panel, revisiting 185 of the first generation VLS households and their split-offs, in addition to 261 newly added households. 

In this study, I collected information on the constraints to educational investment in girls, focusing on marriage related social norms and customs, among the ICRISAT households in three villages: Dokur, Kalman and Shirapur. This research was sponsored by the National Science Foundation and Cornell University in the US. For papers following this research, see my publication list on the University's website. 

Data can be downloaded via FIGSHARE.